Blue Prism_White Paper_ Technical Kickstarter

Knowledge Kickstarter
– Some Common
RPA Questions
What to consider when deploying Enterprise RPA.
blueprism.com

At Blue Prism, we
have been working to
deliver Robotic Process
Automation solutions
across the enterprise
since 2001.
Our product has grown up in the context of
large scale, demanding deployments where
security, resilience and governance are equally,
if not more important, than speed, automation
and simplicity.
This data sheet outlines some common
questions that are posed by customers
who are beginning the RPA journey, and
aims to help organizations to make sense
of RPA technology and to quickly become
knowledgeable enough in the sector to be
able to choose the best approach.
Knowledge Kickstarter -
Some Common RPA Questions
02 | Knowledge Kickstarter – Common RPA Questions
Blue Prism Software Robots - the Virtual Workforce

It may seem odd that to start a discussion
on our product by explaining why we don’t
have specific functionality but let us explain
why we think this is so critical to enterprise
RPA deployments.
A desktop recorder watches how you
navigate the applications on your desktop
and records the steps that you take – it’s a
useful shortcut to building a process quickly
if that’s what is required.
The downside is that by using a recorder, users
are not designing their solution to be re-used
and are making choices based on short term
imperatives rather than long term objectives.
At Blue Prism, we strongly believe that for RPA
to truly deliver value, longevity and resilience,
automations should be carefully planned,
modelled and designed in order to leverage
principles such as re-use, componentization
and economies of scale.
Desktop Recorders make
things easy…don’t they?”
“
Blue Prism has never had a recorder in our software.
And we have no plans to implement one.

How many processes will you have captured?
How will you ensure that you maximise re-use across those processes for ease of
maintenance?
How will you prevent having to record processes over and over again as systems
and processes change?
How intelligent are the scripts that have been recorded in terms of rules and
system responses?
How will the workload be shared, distributed and managed to achieve SLA targets
and accommodate peaks and troughs in demand?
Think about the long term…
Imagine how your environment will look in 6, 12
or 24 months’ time.
Recorded automations include pauses that
comprise fixed numbers of seconds that
break when your systems run slowly. The
scripts will mix business logic and application
logic together. Elements on the screens such
as fields and buttons that are used in the
automations will be duplicated over and
over again, resulting in an inflexible block of
automation that is very difficult to maintain
and change.
This all means that whilst it may be quick
to record a set of “point to point” single
stream scripted automations to run on
desktops of agents, it’s almost impossible to
take these unconnected scripts and deploy
them in an enterprise context, where scale,
availability, manageability and ease of change
management are vital.
Fig 1. A series of unconnected scripts

If a system changes, change the model. Once.
If an application interaction needs to change,
update the object. Once
All of the processes that use this information
inherit these changes, without re-recording
or rebuilding them. And processes can be
written more quickly as the building blocks
that already exist can be re-used in different
business scenarios.
With our approach, the Total Cost of
Ownership actually reduces over time, rather
than increasing.
Blue Prism works differently.
By modelling the parts of the application that you
are going to use first, then creating re-usable objects
that interface with the system like a user, processes
can be layered on top that chain together this library
of objects with business rules to create flexible,
maintainable processes.
Fig 2. Leveraging re-use

Because of this heritage, the architecture of
these software products is not designed to
scale across significant estates or to leverage
centralization and design control – they are
utility products.
One thing to look out for is how an RPA
product captures and then replays the
automation. Many of these software products
produce “scripts” that are compiled into files
that can be run on a desktop.
The disadvantage with this approach is that
these scripts are slow, inefficient and difficult to
maintain. They are also very linear in approach.
In competitive evaluations, we have seen
complex scripts run 3-4 times slower on other
technologies when compared to our object
oriented approach due to this architecture.
Concepts like recursion (calling the same
object from within itself), control logic
(looping through objects and processes) and
conditional processing (call process B based
on the outcome of process A) are not possible
and the linear approach reduces the flexibility
afforded by the ultimate solution.
Scripts, code, objects –
aren’t they all the same thing?”
Up until 2 or 3 years ago, what many RPA vendors
were actually providing were desktop scripting
applications designed to automate short tasks on a
user’s desktop.
“

Think about running the
automations at scale as
well as developing and
maintaining them. Look
at command and control
very carefully when
evaluating a technology.
If an application is designed to be deployed
on an agent’s desktop, but is actually being
deployed as part of a wider RPA initiative,
consider how the automations will be
controlled and administered as the scale of the
solution grows.
What happens when there is a system problem
or the automation path doesn’t match the
data encountered in the application? Will the
process stop? Who is required to restart it, and
how will they do this? How will you ensure
that the administrator can’t see the customer
data, or interfere with it fraudulently in the
name of the robot?
Similarly, think about scale. If you had 1000
robots, how many administrators will you
need to start the processes, log into the
systems, keep them running? Blue Prism
takes care of all of this with our secure
enterprise infrastructure.
It may take a little longer to set up, but the
result is a secure, reliable system that runs
24/7 and only requires one or two people
to monitor (who don’t have access to the
systems, just the control room).
With a solution that has a desktop heritage,
consider how many people will be needed
to maintain and monitor it, and how you
can ensure that your valuable customer
information is safe and secure.
Aren’t all robots the same when
it comes to running them?”
“

How will you deal with cases outside the scope
of your automation? What about a 2-phase
process, or retrying a case? If things go wrong,
how will you know what was processed, what
was left incomplete and what was unfinished?
Workflow is very important
when it comes to scaling
your solution.
Also, think about rolling back an automation,
sharing a file across a team of robots and
prioritising work with visibility and reporting
on volumes, unit times and backlogs.
Without centralised workflow, none of this
is possible.
And workflow doesn’t just mean coding a
schedule for a robot. It manages the state
of the work items, tracks them, records any
exceptions for later review and allows you to
prioritise, retry, analyse and predict completion
times across your automation estate.
A badly designed solution that doesn’t
consider these facets will impact your business
benefit, lengthen your SLAs and may cause
problems with your governance, compliance
and information security teams. Often this
won’t become apparent until you are some
way into the journey.
See the bigger picture –
workflow.

But a robot is a technology, not a person,
and controls can and should be put in place
to ensure that the data the robots are using
stays secure.
This means protecting the RPA infrastructure
from human intervention, locking down
the robots so they cannot be accessed by
developers, testers or administrators.
When you deploy your solution, your robots have
access to data held within your information systems.
They use this data to perform the automated process
as a human would.
It also means ensuring that the data used by
the robots is encrypted during transportation
and whilst at rest, and having robust access
management policies that log any changes and
enforce role based permissions to demonstrate
to auditors that best practise is being followed
and no one individual has the autonomy to
make a change without a secondary individual
approving that change.
This also means that the system should be
centrally controlled, not via individual scripts
or executables, but by a centralised, server
oriented architecture that uses a database
to store all of the artifacts required by the
RPA platform.
And what about your data?”
“

Either the customer is on the phone and the
agent would like to access more information
quickly to help them with the conversation, or
they want to do something in real time, like
producing a quotation to give to the customer.
Desktop automation, mashups or reskinning
applications for real-time, desktop processing
is a very different approach to RPA and
presents challenges and difficulties that can
be hard to overcome.
Sometimes this is called “attended”
automation, versus “unattended” automation.
You can think of Blue Prism as a rapid
response team of virtual workers in a separate
back office. You delegate them work to do,
they do it, and then report back.
This approach means you can use techniques
like workflow, prioritisation and scheduling
to process large volumes of work securely,
freeing up the desktop of your human
agents, and allowing you to better meet
SLAs and workload volumes and creating a
whole “virtual backoffice” of staff to perform
processes that were previously unachievable.
What’s the difference between
‘front office’ and ‘back office’?”
“
We see a lot of requests for process automations
that originate with a front office requirement

Deploying an automation
on an agent’s desktop,
alongside their day to
day activities, introduces
complexities that are
different to the controlled
environment described
above, that can mean
longer development times
and dissatisfaction with
the resultant deliverable.
Some examples of the problems that can be
encountered range from different desktop
settings or user-customised applications –
the robots need to be trained to work with
every combination of system appearance that
the user might have – through to handling
unsuccessful processes – if a process stops
halfway through because the system or
data didn’t match the design, the user is left
talking to the customer with no information
and no feedback – this can impact the user’s
perception of the solution.
Furthermore, having an agent share a screen
with a robot introduces a lack of control
– the agent can interrupt the automation,
close applications down, introduce unknown
variables and even commit fraud in the name
of the robot as there is no central audit trail,
no control and no security in place for
the approach.
This is why Blue Prism only deliver server side,
unattended automations. The virtual workforce
is secure, reliable and scalable, and can’t be
interrupted by human staff as they do not have
access to this new team of virtual workers.
What’s wrong with a
‘front-office’ approach?”
“

One frequently
explored scenario is
that of a process being
performed in real-time
whilst a human agent
in the front office waits
for the response – again
this can be challenging
to resource.
Imagine a scenario where 1000 agents all need
a real-time robot to assist them with some
data collection or system based calculation.
How many might need a robot simultaneously
at peak times? 200? 300? In this scenario,
you might need 300 robots sitting in the data
centre waiting for a real-time request.
This has a number of drawbacks. First of all,
it’s a waste of resource – why have these
machines sat waiting when they could be
delivering value to your organisation.
Secondly, what if 301 agents all asked for the
information at the same time? – you wouldn’t
want to force the agent who came in last to
have to keep the customer talking whilst they
waited for the robot to service their request –
again, perception is key and cultural adoption
is critical to a project’s success.
So can my server side robots
provide information to users
in real time?”
“

Consider what needs to be done in real-time,
and what can be queued for the robotic workers
to process as part of their daily workload.
Hand-offs through e-forms, existing systems
and workflow can be a very effective way of
splitting processing into online and offline
segments to optimise both human and robotic
workforces and get the best from both.
What’s the solution for ‘front
office’ processes then?”
A better approach is
to look at the process
through a different lens.
“

That said, some processes
do need a human decision
or involvement along
the way.
Examples might be an authorisation part way
through the process, or it could be triggered
by a customer conversation or some subjective
reasoning by a human in the call centre.
In these scenarios you can use standard
workflow techniques to achieve this by way of
a hand-off – work can be moved between the
respective workforces just like you would with
existing teams, either as a call wrap function,
or as an authorisation step before pushing the
work back to the robots to complete.
And because Blue Prism includes powerful
workflow in our solution, you can resource
processes appropriately for their SLA’s and deal
with the priority cases first.
How can I automate a
process that needs human
intervention?”
“

1. Utilise Object Models 2. Design libraries of objects 3. Command and Control
NO Recorders NO Scripts NO Intervention
• Desktop recording
cuts corners
• No design required
• Automation is brittle
• It enforces bad discipline
• It becomes impossible
to maintain
• No workflow or optimisation
• Scripts produce linear,
inflexible processes
• No design required
• Re-use is very difficult
• It enforces bad discipline
• It becomes impossible
to maintain
• Slow performance for
complex processes
• Using human intervention
to start, restart and stop
processes is a security risk
• Business cases are hard
to prove
• Solutions are very difficult
to scale
• Workflow is uncontrolled
• Data is not encrypted
or secured
Blue Prism uses an enterprise-centric approach
to RPA:
What does ‘Enterprise
RPA’ mean?”
REC
“

Imagine the Possibility:
Software Robots –
the Virtual Workforce
The more it learns, the more efficient it
becomes. It works without ever taking a
vacation. It can be small one day or large
when your business hits a spike.
Imagine a different kind of workforce that you can
teach countless skills.
And it frees up your best people in both the
line of business and IT to do their best work on
your highest priorities.
Meet the Software Robots – the Virtual Workforce.
Call us today on
1-888-75-PRISM (1 888-757-7476)
or +44 (0) 870 879 3000
or visit us at www.blueprism.com
to start building your Virtual Workforce.
® Blue Prism is a registered trademark of Blue Prism, Ltd. Copyright © 2016 Blue Prism Ltd

Blue Prism_White Paper_ Technical Kickstarter

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Was ist angesagt?

Was ist angesagt? (20)

Andere mochten auch

Andere mochten auch (19)

Ähnlich wie Blue Prism_White Paper_ Technical Kickstarter

Ähnlich wie Blue Prism_White Paper_ Technical Kickstarter (20)

Blue Prism_White Paper_ Technical Kickstarter